Measurement methods, instruments and calibration

Define the quantity you intend to measure

The measurand is the quantity intended to be measured. “Width of the part” may be incomplete if the part tapers or bends. A more useful description is the maximum outside width across the two fitting faces, at room conditions, without squeezing the part. That definition tells an operator where to look and which contact might change the answer.

A method describes the broad approach. A procedure gives enough detail to carry it out: identify the faces, remove loose debris, align the instrument, apply the specified contact, wait for a stable reading, record the value, and repeat as planned. Different tasks need different details. A soft object can compress under contact, and a photograph can distort length unless its scale and viewpoint are controlled.

Match the instrument to the task

Check the measuring range, the smallest change the instrument can resolve, and the uncertainty achieved in this use. A device that displays hundredths of a millimetre does not automatically deliver hundredth-of-a-millimetre uncertainty. Alignment, contact, temperature, surface condition and calibration can matter more than the final displayed digit.

Also consider response time and disturbance of the object. A slow temperature sensor may miss a brief peak; an oversized probe can change the flow it is intended to observe. Some quantities are found indirectly: density, for example, can be calculated from a mass and a volume. The procedure must then include both measurements and the calculation. An instrument is only one part of a measuring system.

Connect indications to reference values

Calibration establishes a relationship between reference quantity values and instrument indications, including associated uncertainties, and uses that relationship to obtain measurement results. A reference standard is not an infinitely exact object: its assigned value has an uncertainty too. A calibration record must specify the conditions and the range in which the relationship is supported.

In the invented gauge example, reference lengths of 0.0 mm, 10.0 mm and 20.0 mm produce mean indications of 0.2 mm, 10.2 mm and 20.2 mm. Plot indication minus reference. All three central differences are +0.2 mm. An additive correction of −0.2 mm is a reasonable teaching model across that checked range; a real evaluation would also examine uncertainty, repeatability and behaviour between the points.

The graph shows central differences only. It does not contain enough information to calculate a calibration uncertainty. Nor do three points prove the gauge is linear everywhere. Extrapolating the correction beyond 20 mm requires evidence.

Keep three operations distinct

Adjustment changes the measuring system, such as altering a setting to reduce an offset. Calibration characterises its relationship to references; it need not change the instrument. Verification checks whether stated requirements are met. After an adjustment, the previous calibration relationship may no longer apply and must be reassessed as the procedure requires.

Calibration intervals depend on use, stability, environment and the consequence of a poor result. A drop, overload, unusual check result or changed setup can justify investigation before the next scheduled date. A current label is useful evidence, but it cannot make every later measurement valid.

Apply a documented correction

Suppose the gauge indicates 12.6 mm under the same conditions as the example calibration. With the adopted offset model, corrected length = indication − 0.2 mm = 12.4 mm. Preserve 12.6 mm as the raw indication and record the correction separately so someone can reconstruct the result.

If the reading were 25.0 mm, subtracting 0.2 mm would be easy arithmetic but unsupported use of the checked range. Select a suitably characterised range or obtain further calibration evidence before relying on the result. Report the applicable uncertainty with a final measurement; none can be derived from these central values alone.

Try the reasoning

1. Why is “maximum width across the fitting faces without squeezing” better than “width”?

   Show the answer and reasoning

   It defines location, purpose and a relevant condition. Without them, two careful operators could measure different quantities.

2. A display has 0.01 mm steps. Does that establish uncertainty of ±0.01 mm?

   Show the answer and reasoning

   No. Display steps describe only one feature. Calibration, contact, alignment, environment and other effects must be evaluated.

3. Must calibration physically change an instrument?

   Show the answer and reasoning

   No. It establishes the relationship between indications and references, with uncertainties. Adjustment is the operation that changes the system.

What to carry forward

Define the measurand first, then choose and document the whole procedure. Calibration supports a result within stated conditions and limits.